AN ORDINAL PATTERN-ANALYSIS OF 4 HYPOTHESES DESCRIBING THE INTERACTIONS BETWEEN DRUG-ADDICTED, CHRONICALLY DISADVANTAGED, AND MIDDLE-CLASS MOTHER-INFANT DYADS

This study investigated mother-infant interactions in 18 dyads. All pa rticipants were African American and enrolled in an early intervention program because the infants (2-26 months of age) had developmental di sabilities or were at high risk for developmental disability. Some mot hers had used drugs during their pregnancy, and all mothers were of lo w or middle socioeconomic status. Dyads were videotaped interacting at 4 different times, separated by at least 5 months in time. Videotapes were rated in terms of infant involvement and maternal responsivity i n the interaction. 4 hypotheses concerning the pattern of maternal int eraction across time were tested using ordinal pattern analysis. The h ypothesis that mothers would become less responsive to infants over ti me (H(D)) as a function of drug addiction, poverty, or serious develop mental delay was supported for only 4 of the 18 dyads. There was suppo rt for the hypothesis (H(I)) that mothers naturally increase their res ponsivity over time (N = 6) and support for the hypothesis (H(T)) that mothers' interactive sensitivity fluctuates in relation to infants' i nvolvement in the interaction over time (N = 7). Ordinal pattern analy sis has advantages in determining how well competing hypotheses descri be individuals within populations relative to approaches that identify differences that apply to entire populations.